
# Libraries
import torch
import glob
import re
import os
import ast
import random
import numpy as np
import pandas as pd
import seaborn as sns
import shutil as sh
from pathlib import Path
import PIL.Image
import PIL.ImageDraw
import matplotlib.pyplot as plt
%matplotlib inline
from tqdm.auto import tqdm
from IPython.display import Image, clear_output


# Clone the yoloV5 repo
#%%time

!git clone https://github.com/ultralytics/yolov5 # clone repo
!pip install -qr yolov5/requirements.txt  # install dependencies
!cp yolov5/requirements.txt ./

Cloning into 'yolov5'...
remote: Enumerating objects: 12390, done.
remote: Counting objects: 100% (7/7), done.
remote: Compressing objects: 100% (6/6), done.
remote: Total 12390 (delta 1), reused 7 (delta 1), pack-reused 12383
Receiving objects: 100% (12390/12390), 11.51 MiB | 11.68 MiB/s, done.
Resolving deltas: 100% (8622/8622), done.


# Location of dataset
base_path = Path('oil_storage_data')
base_path_str = 'oil_storage_data'

# List all images in the folder
image_list = list(base_path.glob('images/*.jpg'))

# Create a list of image ids
image_ids = pd.DataFrame(image_list).rename(columns={0:"image_id"})

# Show a random image
pickone = random.choice(image_list)
img = PIL.Image.open(pickone)
filename = "oil-storage-sample.jpg"
img.save(filename)
display(img)


only_files = [base_path / f.name for f in image_list if os.path.isfile(f) and f.name[-4:] == ".jpg"]
print("Found {} images files in {}".format(len(only_files), base_path))

image_height, image_width = img.size
num_channels = len(img.mode)
print("Image size: {}".format((image_height, image_width)))
print("Num channels: {}".format(num_channels))

# Image is 2560x2560 with 3 color channels

Found 98 images files in oil_storage_data
Image size: (2560, 2560)
Num channels: 3


df = pd.read_csv(base_path / 'annotations.csv')
# convert a string record into a valid python object
def f(x): 
    return ast.literal_eval(x.rstrip('\r\n'))

df = pd.read_csv(base_path / "annotations.csv", 
                converters={'bounds': f})

df["image_id"] = df["image_id"] + ".jpg"

df.rename(columns = {"bounds": "geometry"}, inplace = True)

labels = pd.read_csv(base_path / "annotations.csv", 
                converters={'bounds': f})

df.head(10)


# Create polygon from bounds
def create_polygon_from_bounds(bbox):
    # Get the x and y min and maxes
    (xmin, ymin, xmax, ymax) = bbox
    # Make a coordinates list
    coords = []
    coords.append((xmin, ymin))
    coords.append((xmin, ymax))
    coords.append((xmax, ymax))
    coords.append((xmax, ymin))
    coords.append((xmin, ymin))
    return coords

def overlay_image(image_id, bbox_df):
    img = PIL.Image.open(base_path_str + "/images/" + str(image_id) + ".jpg")
    draw = PIL.ImageDraw.Draw(img)

    for k, row in bbox_df[bbox_df['image_id'] == image_id].iterrows():
        geometry = create_polygon_from_bounds(row['bounds'])
        draw.polygon(geometry, outline=(255,0,0), width=2)
        #draw.text(geometry[0], row['class'], fill=(255,0,0))
        
    return img

# Convert the previous pickone to a string, striping everying but ID
pickone = str(pickone).lstrip(f"{base_path}/images/").rstrip(".jpg")

# Show image overlayed with bounding boxes
img = overlay_image(pickone, labels)
filename = "oil-storage-sample_annotated.jpg"
img.save(filename)
display(img)


# create a list of images used for validation
fold = 1
num_fold = 5
index = df['image_id'].unique()
val_indexes = index[len(index)*fold//num_fold:len(index)*(fold+1)//num_fold]
print(val_indexes)

['400cbcc8-6b72-433b-ac46-92bef8ee6c66.jpg'
 'e1a1bf76-a58c-4582-ace2-7fa29ea66200.jpg'
 '5853d123-e81e-401a-a8c7-37b8d2248000.jpg'
 'c1c959c0-6a0c-497b-82c7-129e076ae192.jpg'
 '183a6155-c7ab-41cd-ab0c-9fc500621ae7.jpg'
 'b64ce3b3-251a-42e4-9439-fb068512983c.jpg'
 '3e040964-7583-401a-8ecb-19fccdc9d8a2.jpg'
 '28483ab7-42f0-4343-af24-f8076dbdf2f9.jpg'
 '9892f3a0-f541-43b8-bc62-d640701841f7.jpg'
 'a2c80dde-a31e-4bb7-bac4-af3aa3d14ff7.jpg'
 '76a1f831-9193-47b7-a576-30e846210d5b.jpg'
 'b78898b5-1941-4cbc-8231-5c966c7f13ef.jpg'
 '664e361b-973a-409e-9df7-dc9513a53583.jpg'
 '22b8b29e-43de-4cf5-b91a-2ff5a188c2c5.jpg'
 '5529501e-5b67-4a9f-bedd-efb785b78ce2.jpg'
 'fb0d4981-0106-47a7-acec-c1bb0e0f8c87.jpg'
 'b1f41adb-2ca9-41d4-aaad-6e3ccf6c7701.jpg'
 '8e40cbf3-aea6-497c-83ed-5f9929e4ccde.jpg'
 '99831d12-b531-4da7-925e-3d25f946d1d0.jpg'
 'c61e23c7-f20f-425e-8380-be73772b69fd.jpg']


import tqdm.notebook

def getWidth(bounds):
    try: 
        (xmin, ymin, xmax, ymax) = bounds
        return np.abs(xmax - xmin)
    except:
        return np.nan

def getHeight(bounds):
    try: 
        (xmin, ymin, xmax, ymax) = bounds
        return np.abs(ymax - ymin)
    except:
        return np.nan

# Create bounds, width and height
df.loc[:,'width'] = df.loc[:,'geometry'].apply(getWidth)
df.loc[:,'height'] = df.loc[:,'geometry'].apply(getHeight)
df.head(10)


# Since image is 2560x2560 we need to crop them into tiles. Here I choose to crop into 512x512 with 64 pixel overlap

tile_width = 512
tile_height = 512
tile_overlap = 64
truncated_percent = 0.3
_overwriteFiles = True

# For using YoloV5 we need a specific folder structure for the cropped images
tiles_directories = {"train": Path("yolov5/train/images"),
                    'val': Path('yolov5/val/images/')}

for _, folder in tiles_directories.items():
    if not os.path.isdir(folder):
        os.makedirs(folder)
        
labels_directory = {"train": Path("yolov5/train/labels"),
                   'val': Path('yolov5/val/labels/')}

for _, folder in labels_directory.items():
    if not os.path.isdir(folder):
        os.makedirs(folder)


# Save one line in .txt file for each tag found inside the tile
def tag_is_inside_tile(bounds, x_start, y_start, width, height, truncated_percent):
    x_min, y_min, x_max, y_max = bounds
    x_min, y_min, x_max, y_max = x_min - x_start, y_min - y_start, x_max - x_start, y_max - y_start

    if (x_min > width) or (x_max < 0.0) or (y_min > height) or (y_max < 0.0):
        return None
    
    x_max_trunc = min(x_max, width) 
    x_min_trunc = max(x_min, 0) 
    if (x_max_trunc - x_min_trunc) / (x_max - x_min) < truncated_percent:
        return None

    y_max_trunc = min(y_max, width) 
    y_min_trunc = max(y_min, 0) 
    if (y_max_trunc - y_min_trunc) / (y_max - y_min) < truncated_percent:
        return None
        
    x_center = (x_min_trunc + x_max_trunc) / 2.0 / width
    y_center = (y_min_trunc + y_max_trunc) / 2.0 / height
    x_extend = (x_max_trunc - x_min_trunc) / width
    y_extend = (y_max_trunc - y_min_trunc) / height
    
    return (0, x_center, y_center, x_extend, y_extend)


# Crop the images into 512x512 and store in YoloV5 compatible folders
for img_path in tqdm.notebook.tqdm(image_list):
    # Open image and related data
    pil_img = PIL.Image.open(img_path, mode='r')
    np_img = np.array(pil_img, dtype=np.uint8)

    # Get annotations for image
    img_labels = df[df["image_id"] == img_path.name]

    # Count number of sections to make
    X_tiles = (image_width + tile_width + tile_overlap - 1) // tile_width 
    Y_tiles = (image_height + tile_height + tile_overlap - 1) // tile_height
    
    # Cut each tile
    for x in range(X_tiles):
        for y in range(Y_tiles):

            x_end = min((x + 1) * tile_width - tile_overlap * (x != 0), image_width)
            x_start = x_end - tile_width
            y_end = min((y + 1) * image_height - tile_overlap * (y != 0), image_height)
            y_start = y_end - tile_height
            
            folder = 'val' if img_path.name in val_indexes else 'train'
            save_tile_path = tiles_directories[folder].joinpath(img_path.stem + "_" + str(x_start) + "_" + str(y_start) + ".jpg")
            save_label_path = labels_directory[folder].joinpath(img_path.stem + "_" + str(x_start) + "_" + str(y_start) + ".txt")
                
            # Save if file doesn't exit
            if _overwriteFiles or not os.path.isfile(save_tile_path):
                cut_tile = np.zeros(shape=(tile_width, tile_height, 3), dtype=np.uint8)
                cut_tile[0:tile_height, 0:tile_width, :] = np_img[y_start:y_end, x_start:x_end, :]
                cut_tile_img = PIL.Image.fromarray(cut_tile)
                cut_tile_img.save(save_tile_path)

            found_tags = [
                tag_is_inside_tile(bounds, x_start, y_start, tile_width, tile_height, truncated_percent)
                for i, bounds in enumerate(img_labels['geometry'])]
            found_tags = [el for el in found_tags if el is not None]

            # save labels
            with open(save_label_path, 'w+') as f:
                for tags in found_tags:
                    f.write(' '.join(str(x) for x in tags) + '\n')

  0%|          | 0/98 [00:00<?, ?it/s]


# See how many training images we have after cropping
n_train = len(os.listdir("yolov5/train/images"))
n_val = len(os.listdir("yolov5/val/images")) 
print(f"There are {n_train} training images after cropping")
print(f"There are {n_val} validation images after cropping")

There are 468 training images after cropping
There are 120 validation images after cropping


# Tensorboard (optional)
#%load_ext tensorboard
#%tensorboard --logdir runs/train


# Create the yaml file required by YoloV5
CONFIG = """
# train and val datasets (image directory or *.txt file with image paths)
train: train/
val: val/

# number of classes
nc: 1

# class names
names: ['oil-storage-tank']
"""

with open("Oil_Storage.yaml", "w") as f:
    f.write(CONFIG)


print('Setup complete. Using torch %s %s' % (torch.__version__, torch.cuda.get_device_properties(0) if torch.cuda.is_available() else 'CPU'))

Setup complete. Using torch 1.11.0+cu102 _CudaDeviceProperties(name='NVIDIA GeForce RTX 2080 SUPER', major=7, minor=5, total_memory=7959MB, multi_processor_count=48)


!python yolov5/train.py --img 512 --batch 32 --epochs 50 --data Oil_Storage.yaml --weights yolov5s.pt

train: weights=yolov5s.pt, cfg=, data=Oil_Storage.yaml, hyp=yolov5/data/hyps/hyp.scratch-low.yaml, epochs=50, batch_size=32, imgsz=512, rect=False, resume=False, nosave=False, noval=False, noautoanchor=False, evolve=None, bucket=, cache=None, image_weights=False, device=, multi_scale=False, single_cls=False, optimizer=SGD, sync_bn=False, workers=8, project=yolov5/runs/train, name=exp, exist_ok=False, quad=False, cos_lr=False, label_smoothing=0.0, patience=100, freeze=[0], save_period=-1, local_rank=-1, entity=None, upload_dataset=False, bbox_interval=-1, artifact_alias=latest
github: up to date with https://github.com/ultralytics/yolov5 ✅
YOLOv5 🚀 v6.1-82-g71621df torch 1.11.0+cu102 CUDA:0 (NVIDIA GeForce RTX 2080 SUPER, 7960MiB)

hyperparameters: lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=0.05, cls=0.5, cls_pw=1.0, obj=1.0, obj_pw=1.0, iou_t=0.2, anchor_t=4.0, fl_gamma=0.0, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, mosaic=1.0, mixup=0.0, copy_paste=0.0
Weights & Biases: run 'pip install wandb' to automatically track and visualize YOLOv5 🚀 runs (RECOMMENDED)
TensorBoard: Start with 'tensorboard --logdir yolov5/runs/train', view at http://localhost:6006/
Downloading https://github.com/ultralytics/yolov5/releases/download/v6.1/yolov5s.pt to yolov5s.pt...
100%|██████████████████████████████████████| 14.1M/14.1M [00:01<00:00, 13.3MB/s]

Overriding model.yaml nc=80 with nc=1

                 from  n    params  module                                  arguments                     
  0                -1  1      3520  models.common.Conv                      [3, 32, 6, 2, 2]              
  1                -1  1     18560  models.common.Conv                      [32, 64, 3, 2]                
  2                -1  1     18816  models.common.C3                        [64, 64, 1]                   
  3                -1  1     73984  models.common.Conv                      [64, 128, 3, 2]               
  4                -1  2    115712  models.common.C3                        [128, 128, 2]                 
  5                -1  1    295424  models.common.Conv                      [128, 256, 3, 2]              
  6                -1  3    625152  models.common.C3                        [256, 256, 3]                 
  7                -1  1   1180672  models.common.Conv                      [256, 512, 3, 2]              
  8                -1  1   1182720  models.common.C3                        [512, 512, 1]                 
  9                -1  1    656896  models.common.SPPF                      [512, 512, 5]                 
 10                -1  1    131584  models.common.Conv                      [512, 256, 1, 1]              
 11                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          
 12           [-1, 6]  1         0  models.common.Concat                    [1]                           
 13                -1  1    361984  models.common.C3                        [512, 256, 1, False]          
 14                -1  1     33024  models.common.Conv                      [256, 128, 1, 1]              
 15                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          
 16           [-1, 4]  1         0  models.common.Concat                    [1]                           
 17                -1  1     90880  models.common.C3                        [256, 128, 1, False]          
 18                -1  1    147712  models.common.Conv                      [128, 128, 3, 2]              
 19          [-1, 14]  1         0  models.common.Concat                    [1]                           
 20                -1  1    296448  models.common.C3                        [256, 256, 1, False]          
 21                -1  1    590336  models.common.Conv                      [256, 256, 3, 2]              
 22          [-1, 10]  1         0  models.common.Concat                    [1]                           
 23                -1  1   1182720  models.common.C3                        [512, 512, 1, False]          
 24      [17, 20, 23]  1     16182  models.yolo.Detect                      [1, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]]
Model summary: 270 layers, 7022326 parameters, 7022326 gradients, 15.8 GFLOPs

Transferred 343/349 items from yolov5s.pt
Scaled weight_decay = 0.0005
optimizer: SGD with parameter groups 57 weight (no decay), 60 weight, 60 bias
train: Scanning '/home/theo/Documents/oil_storage-detector/yolov5/train/labels' 
train: New cache created: /home/theo/Documents/oil_storage-detector/yolov5/train/labels.cache
val: Scanning '/home/theo/Documents/oil_storage-detector/yolov5/val/labels' imag
val: New cache created: /home/theo/Documents/oil_storage-detector/yolov5/val/labels.cache
Plotting labels to yolov5/runs/train/exp/labels.jpg... 

AutoAnchor: 4.42 anchors/target, 0.999 Best Possible Recall (BPR). Current anchors are a good fit to dataset ✅
Image sizes 512 train, 512 val
Using 8 dataloader workers
Logging results to yolov5/runs/train/exp
Starting training for 50 epochs...

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      0/49     4.15G    0.1249   0.01959         0        50       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297   0.000542    0.00337   0.000103   1.91e-05

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      1/49     5.52G    0.1088   0.01861         0        16       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297     0.0018     0.0539    0.00066   0.000196

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      2/49     5.52G   0.09181   0.02277         0        43       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.257     0.0741     0.0365     0.0082

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      3/49     5.52G   0.08184   0.01642         0        51       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297        0.2      0.246        0.1     0.0186

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      4/49     5.52G    0.0811   0.01938         0        34       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.417       0.31      0.243     0.0623

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      5/49     5.52G   0.07654   0.01554         0        64       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.515      0.391      0.363       0.13

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      6/49     5.52G   0.07201   0.01504         0       102       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.321      0.323      0.161     0.0463

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      7/49     5.52G   0.06382   0.01679         0        48       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.563      0.343      0.337      0.113

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      8/49     5.52G    0.0572   0.01266         0        21       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.476      0.438      0.344      0.131

     Epoch   gpu_mem       box       obj       cls    labels  img_size
      9/49     5.52G   0.05942   0.01218         0        17       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.694      0.413      0.488      0.234

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     10/49     5.52G   0.05606   0.01149         0        68       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.726      0.448      0.479      0.257

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     11/49     5.52G   0.05422    0.0131         0       104       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297       0.71      0.545      0.587       0.27

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     12/49     5.52G   0.05202   0.01193         0        48       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.714      0.545       0.51      0.226

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     13/49     5.52G   0.04823    0.0133         0        36       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.792      0.589      0.639      0.322

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     14/49     5.52G   0.05237   0.01037         0       126       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.776      0.519      0.587      0.258

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     15/49     5.52G   0.05204  0.009836         0        26       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.811      0.579      0.654      0.322

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     16/49     5.52G   0.04804  0.009452         0        67       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.895      0.552      0.664      0.372

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     17/49     5.52G   0.04955   0.00816         0        75       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.904      0.599      0.684      0.373

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     18/49     5.52G   0.04695  0.009371         0        76       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.864      0.643      0.703      0.399

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     19/49     5.52G   0.04883   0.01076         0        41       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.845      0.586      0.663      0.358

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     20/49     5.52G   0.04426  0.008695         0        66       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.915      0.582      0.674      0.397

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     21/49     5.52G   0.04124   0.01004         0        15       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297       0.92      0.579      0.679      0.363

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     22/49     5.52G   0.04684   0.01037         0        13       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.902      0.623      0.709      0.368

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     23/49     5.52G   0.04678  0.008734         0        35       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.883      0.636      0.703      0.418

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     24/49     5.52G   0.04249  0.008631         0        94       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.868      0.603      0.681      0.405

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     25/49     5.52G    0.0431  0.008705         0       101       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.897      0.643      0.724      0.429

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     26/49     5.52G   0.03972  0.007632         0        37       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.891      0.613      0.695      0.401

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     27/49     5.52G   0.04352  0.008978         0       194       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.866      0.653      0.738      0.438

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     28/49     5.52G   0.03928  0.007462         0        28       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.836      0.616      0.723      0.343

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     29/49     5.52G   0.04468  0.007272         0        69       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.782      0.663      0.717      0.423

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     30/49     5.52G   0.03864  0.006491         0        31       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.883      0.609      0.709      0.401

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     31/49     5.52G   0.03721   0.00846         0        52       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.846      0.609      0.686       0.38

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     32/49     5.52G   0.03655  0.007261         0        58       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.922      0.606      0.706       0.44

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     33/49     5.52G   0.03669  0.006385         0        74       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.865      0.606      0.679      0.431

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     34/49     5.52G   0.03509  0.007559         0        15       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.874      0.562      0.655      0.414

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     35/49     5.52G   0.03615  0.007207         0        27       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.861      0.542      0.649      0.411

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     36/49     5.52G   0.03501  0.006353         0        11       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297       0.84      0.603      0.697      0.426

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     37/49     5.52G   0.03797  0.007889         0        53       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.935      0.579      0.712      0.429

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     38/49     5.52G   0.03576  0.005833         0        19       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.892      0.636      0.719      0.443

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     39/49     5.52G   0.03548  0.007155         0        71       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.891      0.687      0.745      0.455

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     40/49     5.52G   0.03706  0.006631         0        19       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.895      0.663       0.73      0.463

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     41/49     5.52G   0.03234  0.007324         0        45       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.829       0.67      0.728      0.464

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     42/49     5.52G   0.03289  0.007524         0        33       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.825      0.684      0.732      0.456

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     43/49     5.52G    0.0323  0.007006         0         4       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.829      0.653      0.716      0.445

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     44/49     5.52G   0.03144  0.005577         0        13       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.864      0.643      0.714       0.45

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     45/49     5.52G    0.0318  0.007072         0        25       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.872      0.663      0.725      0.454

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     46/49     5.52G   0.03237  0.007188         0        39       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.885       0.65      0.738      0.455

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     47/49     5.52G    0.0296  0.006488         0        41       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.889      0.657      0.739      0.459

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     48/49     5.52G   0.03273  0.006672         0        66       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.876      0.643       0.73      0.457

     Epoch   gpu_mem       box       obj       cls    labels  img_size
     49/49     5.52G   0.03008  0.006422         0         9       512: 100%|███
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.822       0.69      0.733      0.461

50 epochs completed in 0.055 hours.
Optimizer stripped from yolov5/runs/train/exp/weights/last.pt, 14.4MB
Optimizer stripped from yolov5/runs/train/exp/weights/best.pt, 14.4MB

Validating yolov5/runs/train/exp/weights/best.pt...
Fusing layers... 
Model summary: 213 layers, 7012822 parameters, 0 gradients, 15.8 GFLOPs
               Class     Images     Labels          P          R     mAP@.5 mAP@
                 all        120        297      0.829       0.67      0.728      0.464
Results saved to yolov5/runs/train/exp
Exception ignored in: <function _MultiProcessingDataLoaderIter.__del__ at 0x7f70eb69d3b0>
Traceback (most recent call last):
  File "/home/theo/anaconda3/envs/oil_dl_env/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 1358, in __del__
  File "/home/theo/anaconda3/envs/oil_dl_env/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 1283, in _shutdown_workers
AttributeError: 'NoneType' object has no attribute 'python_exit_status'
Exception ignored in: <function _MultiProcessingDataLoaderIter.__del__ at 0x7f70eb69d3b0>
Traceback (most recent call last):
  File "/home/theo/anaconda3/envs/oil_dl_env/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 1358, in __del__
  File "/home/theo/anaconda3/envs/oil_dl_env/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 1283, in _shutdown_workers
AttributeError: 'NoneType' object has no attribute 'python_exit_status'


!python yolov5/detect.py --source oil_storage_data/extras --img-size 2560 --weights yolov5/runs/train/exp/weights/best.pt --conf 0.5

detect: weights=['yolov5/runs/train/exp/weights/best.pt'], source=oil_storage_data/extras, data=yolov5/data/coco128.yaml, imgsz=[2560, 2560], conf_thres=0.5, iou_thres=0.45, max_det=1000, device=, view_img=False, save_txt=False, save_conf=False, save_crop=False, nosave=False, classes=None, agnostic_nms=False, augment=False, visualize=False, update=False, project=yolov5/runs/detect, name=exp, exist_ok=False, line_thickness=3, hide_labels=False, hide_conf=False, half=False, dnn=False
YOLOv5 🚀 v6.1-82-g71621df torch 1.11.0+cu102 CUDA:0 (NVIDIA GeForce RTX 2080 SUPER, 7960MiB)

Fusing layers... 
Model summary: 213 layers, 7012822 parameters, 0 gradients, 15.8 GFLOPs
image 1/5 /home/theo/Documents/oil_storage-detector/oil_storage_data/extras/588fc1fb-b86a-4fb4-8161-d9bd3a1556ca.jpg: 2560x2560 134 oil-storage-tanks, Done. (0.053s)
image 2/5 /home/theo/Documents/oil_storage-detector/oil_storage_data/extras/605ffac0-69d5-4748-92c2-48d43f51afc1.jpg: 2560x2560 132 oil-storage-tanks, Done. (0.049s)
image 3/5 /home/theo/Documents/oil_storage-detector/oil_storage_data/extras/67f7c7ad-11a1-4c7f-9f2a-da7ef50bfdd8.jpg: 2560x2560 26 oil-storage-tanks, Done. (0.049s)
image 4/5 /home/theo/Documents/oil_storage-detector/oil_storage_data/extras/b8c0e212-3669-4ff8-81a5-32191d456f86.jpg: 2560x2560 89 oil-storage-tanks, Done. (0.048s)
image 5/5 /home/theo/Documents/oil_storage-detector/oil_storage_data/extras/df5ec618-c1f3-4cfe-88b1-86799d23c22d.jpg: 2560x2560 88 oil-storage-tanks, Done. (0.048s)
Speed: 2.7ms pre-process, 49.5ms inference, 0.7ms NMS per image at shape (1, 3, 2560, 2560)
Results saved to yolov5/runs/detect/exp

	image_id	class	geometry
0	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(1205, 1493, 1231, 1516)
1	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(1479, 1756, 1507, 1776)
2	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(1478, 1786, 1518, 1815)
3	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(2132, 2163, 2148, 2178)
4	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(2010, 2127, 2024, 2141)
5	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(1991, 2130, 2009, 2150)
6	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(1971, 2139, 1992, 2158)
7	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(2099, 2004, 2108, 2013)
8	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(2108, 2005, 2115, 2012)
9	0918a15c-8cfb-4da7-92cf-1f5098a76760.jpg	oil-storage-tank	(2103, 1645, 2113, 1656)

Oil-Storage Detector¶